Calculate edge in sports betting by subtracting the implied probability of the price from your own fair probability. Positive means you are being paid for a side you think wins more often than the odds allow — negative means the house is charging you a vig (a booking fee) to take the other side. The honest part of the math is the benchmark, and our live feed proves it: on August 24 every matchup posted one book with zero dispersion, so you can see exactly when there is no fair number to compare against.
How do you calculate edge in sports betting?
Your edge is your fair probability minus the implied probability in the price, and the whole job is getting that fair probability right. Convert the offered odds to a percentage (implied probability), run your model for its own probability (fair probability), then subtract. A positive result is an edge; a negative one is a tax. The formula is trivial — the hard part is the two inputs, and our feed shows why.
Where the price comes from: implied probability
Every line carries a probability inside it. The bigger the favorite, the fewer times it must win to break even, and the smaller the underdog's implied shot. The vig (the built-in booking fee) means the two sides of any game add up past 100 percent — that extra is what the house keeps.
Look at the Rays at home at spread 1.5 across 1 book . The number itself is a price. Without the vig removed, you cannot say what the fair probability is — and with one book you have no second price to check against. The guide to vig, juice, and hold walks through stripping the fee out of the percentage.
Where your number comes from: fair probability
Your fair probability is what your model actually believes. It can come from a regression, a rating system, or a backtested trend. The best sports-betting models explain the edge math with real features — see our explainer on how spreads work and the closing line value walkthrough for the frameworks.
The gap between fair and implied probability is the whole game. That gap is only useful if your fair number is better than a guess, which is why the market's own opinion matters.
What one book and zero dispersion actually mean
August 24 gives you a clean proof of the failure mode. Tampa Bay Rays at home spread 1.5 across 1 book . Boston Red Sox at home spread 1.5 across 1 book . Colorado Rockies at home spread -1.5 across 1 book . Texas Rangers at home spread -1.5 across 1 book .
One price, zero dispersion (no gap between the highest and lowest line) means the market has shown up exactly once. You can run the edge formula until the numbers bleed, but nobody has vetted your fair probability against a second book. The way to sanity-check it is to log every edge bet against the closing line — that is the track record and the CLV feed in action. When your edge beats the closing line over a real sample, the math is working.
The college blowouts are the cautionary tale
When the spread is enormous, the formula gets dishonest fast. Arkansas-Pine Bluff at Missouri at home spread -54.5 across 1 book . Eastern Illinois at Minnesota at home spread -43.5 across 1 book . Bethune-Cookman at UCF at home spread -42.5 across 1 book . North Carolina at TCU shows the same game twice at -7.5 and -8 .
A half-point hook on a -54.5 game may not move your edge by much, but a half-point on a game near a key number changes everything. The bigger the favorite, the less the raw spread carries about the actual cover probability — so check the tracking discipline before you count a blowout as an edge. Real edges live where the market is competitive, not where it is one price on a mismatch.
What would change our mind
A real number flips this take: if a game currently sitting at one book splits across multiple books and dispersion stays under the spread itself, the price is confirmed and our edge calculation gets trustworthy. North Carolina at TCU at -7.5 and -8 is the closest thing the slate has to a second opinion. Wake me when the feed posts two real books on a close game, and the spread math earns its keep at a price worth your bankroll.
Bet responsibly — set limits, never chase losses. The receipts stay in The Receipts Drawer.
NFL ATS cover-margin distribution
Distribution of (final margin − closing spread) across an NFL season. Roughly normal with mean ≈ 0 and standard deviation ≈ 13 points, which is why most ATS edges live in the ±1.5 point window. Chart ELO standings winner: free-llm-a on team-summary::graded-record (ELO 1516, artifacts/chart-elo/standings.json).
EV per $100 across win rate × odds grid
Expected value of a $100 stake at each combination of true win rate and market odds. Anywhere the cell is positive you have a long-run profitable bet; the magnitude shows how aggressive Kelly will size it.
